Steering vectors from frozen LM layers enable a lightweight classifier to detect machine-generated text robustly across domains, source models, and editing attacks.
Wong and Shu Yang and Xinyi Yang and Yulin Yuan and Lidia S
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SV-Detect: AI-generated Text Detection with Steering Vectors
Steering vectors from frozen LM layers enable a lightweight classifier to detect machine-generated text robustly across domains, source models, and editing attacks.